Learn OpenChainGraph

Everything that explains the standard, in one place

OpenChainGraph turns a decision tool's output into a hash-anchored, agent-callable, verifiable artifact. This hub links every explainer, the technical specification and white paper, the compute-integrity deep dive, and the integration guides.

Written for engineers, auditors, and teams evaluating verifiable decision provenance. New here? Start with the first row and work down.

01 · Start here

Orientation

What the standard is, why it exists, and what it looks like in motion.

Positioning
Why OpenChainGraph
An honest prior-art comparison to in-toto, SLSA, OpenLineage, and AP2: what OCG borrows, what is genuinely new, and why the pieces only converged now.
Read → Interactive
How It Works
A twenty-two-stage walkthrough with a live SHA-256 demo: how artifacts hash, chain by parent_hash, and get verified by agents over MCP.
Open explainer → Comparison
Two Worlds
The compliance audit side by side: email chains, PDFs, and version drift today, versus hash-linked artifacts with OpenChainGraph.
Compare →
02 · Technical reference

The standard, in depth

The normative material, the paper, and the deep dive on proving a computation ran the way it claims.

Normative
Technical Specification
The envelope schema, canonical hashing, the export and proof profiles, and the machine-checked conformance gate suite.
Read the spec → White paper
The Paper
Browser-native, hash-anchored provenance for deterministic, agent-callable decision artifacts, with the formal guarantees and measured costs.
Read the paper → Compute-integrity
Proving It Ran (§18)
Step by step, how a zkVM proof lets a verifier confirm a calculation ran the way it claims, without re-running it or seeing the inputs.
Read the deep dive → Interactive
The AIN Bridge
How browser tools compose: prefill deep-links, Policy Mandate intake, and the postMessage bridge that hands an artifact from one tool to the next.
Open explainer → Deep dive
Machines Decide, Humans See Exceptions
How a gated chain runs the routine case straight through and hands the exception case to a person, with mortgage preflight and adverse action as shipped evidence.
Read the deep dive →
03 · Build and verify

Hands-on tools and integration

Compose, verify, export, and self-certify. Several of these run entirely in your browser.

Integration
Integration Guide
How to implement, self-certify, and badge a tool as OpenChainGraph-conformant. Envelope schema, canonicalization, and the conformance suite.
Read the guide → Export profiles
Export a Verified Artifact
Render a verified artifact as xlsx, csv, pdf, or xbrl with the export profiles, all downstream of a fixed execution_hash.
Read the guide → Educational tool
Chain Builder
A hands-on teaching tool: click to compose a multi-step workflow and watch the parent_hash provenance form as you go.
Try it → Tamper lab
Live Sandbox
Edit an artifact live and watch the execution_hash recompute in your browser. Tamper with a field and see exactly why the fingerprint breaks.
Open sandbox → Verifier
Verify an Artifact
Paste any OpenChainGraph artifact and verify its execution_hash client-side. Nothing leaves your browser.
Verify now → Agent ledger
Agent Work Ledger
Verify agent-produced records in your browser: replay gate decisions, inspect the decision path, and check the execution hash. Accepts dropped files, pasted JSON, or fragment links from MCP tool responses.
Open ledger → Concepts
Where It Fits
Application concepts across generative art, gaming, AI agents, payments, and more, for teams exploring where verifiable decisions belong.
Explore → Live implementation
ApexLogics OpenChainGraph Suite
A sister site running 13 persona-based OpenChainGraphs across 145 tools in production, each step a hash-anchored Policy Mandate artifact.
Open suite →
04 · Standards building blocks

How OCG meets the standards you already use

One short guide per interoperability layer. Each shows the exact mapping and how to declare conformance.

SLSA
buildType and SLSA
Align the buildType URI with SLSA provenance: identify the hash construction and satisfy SLSA levels.
Read → in-toto
in-toto Envelopes
Wrap artifacts in DSSE in-toto envelopes, sign with Ed25519, and compose with Sigstore and cosign.
Read → W3C PROV
W3C PROV-DM
Map OpenChainGraph chains onto the W3C PROV data model for interchange with provenance tooling.
Read → Signing
Ed25519 Keys
Generate, store, and rotate Ed25519 keys for the audit_signature envelope, signing in the browser with WebCrypto.
Read → Conformance
Profile Conformance
Declare profile conformance with dereferenceable dct:conformsTo URIs a verifier can follow.
Read → Payments
ISO 20022
Overlay ISO 20022-aligned payment fields on an artifact for financial-message interchange.
Read → Knowledge
Open Knowledge Format
Add an OKF companion bundle that maps graph edges to machine-readable concept documents.
Read → Observability
OpenTelemetry
Map decision chains onto OpenTelemetry traces: execution_hash to spanId, parent to parentSpanId.
Read → Compliance
Decision-Receipt Crosswalk
Field-by-field alignment: OCG artifact (execution_hash §4, kernel identity §17, anchor_bindings §20) vs. EU AI Act Art 12(2), ECOA Reg B §1002.12, ASC 815, and IETF agent audit trail.
Read →
05 · Autonity Historical Series

How a live decentralised derivatives protocol worked

Five interactive explainers preserving the mechanics of Autonity's Autonity Financial Protocol (AFP) and Decentralised Clearing Contract (DCC). Autonity stood down in February 2026. These pages document what it built — and where every mechanic lives today.

E-01 · Markets
Prediction Market Mechanics
Binary versus scalar forecast contracts, payoff formulas, and an interactive PnL calculator seeded with Autonity Season 1 context.
Open explainer → E-02 · Margin
Margin, Maintenance & Liquidation
How the DCC tracked position health in real time: unrealised PnL, margin buffer, liquidation price, and the GREEN/AMBER/RED health system.
Open explainer → E-03 · Clearing · Centerpiece
AFP/DCC Venue-Agnostic Clearing
The two-key model that separated margin custody from trade intent — and why every mechanic it proved out now runs in production somewhere under a different name.
Open explainer → E-04 · Competition
Forecastathon Season Scoring S1–S4
How Autonity scored its forecasting competition: a four-component composite (performance, volume, accuracy, referrals) for S1–S3, simplified to PnL for S4.
Open explainer → E-05 · Monetary Design
ACU Currency Basket: SDR-Method Valuation
How the Auton Currency Unit tracked a fixed-amount basket of seven currencies using the same methodology the IMF uses for the Special Drawing Right.
Open explainer →